Building Systems That Let People Work Better: An Interview with Vybe Founding Engineer Paul Salmon

“It was never the code that slowed teams down,” he says. “It was everything around it - communication, coordination, trust.”

We are speaking from different time zones, a fitting detail for someone whose career has unfolded across France, the UK, Australia, and the United States. Over the past decade, Salmon has quietly built a reputation as a senior full-stack engineer who gravitates toward the particularly challenging layer of modern technology: the intersection of AI systems, human behavior, and organisational scale. His work spans large enterprises, venture-backed startups, and founder-led experiments, but the through-line is consistent. His work focuses on systems intended to reduce operational friction  - especially in environments where conversation, coaching, and decision-making are core economic activities.

That focus aligns with a broader shift occurring across the global economy. . According to McKinsey Global Institute, up to 30 percent of hours worked globally could be automated by 2030, not by replacing workers outright, but by reshaping how knowledge work is done. The demand, McKinsey notes, is increasingly for tools that augment human capability rather than eliminate it.

“That’s the part people underestimate,” Salmon says. “Automation doesn’t kill work. Bad automation does.”

From Engineering Output to Human Systems

Salmon’s early career followed a familiar technical path: formal engineering training in France, followed by roles at companies such as Thales and Orange, where he learned how large-scale systems behave under real-world constraints. But it was his move into startups - and later into AI-driven products - that reframed how he understood impact.

While working in Sydney on early natural-language processing systems, he helped build pipelines that processed more than 100,000 news articles per day. The challenge was not simply computational scale; it was relevance.

“You can summarise anything,” he explains. “The hard part is knowing what matters to a human on the other side.”

That lesson followed him to San Francisco, where he became Lead Engineer at Plato, a mentorship platform designed for senior engineering leaders. The company progressed from an early-stage product to a Series A-backed organisation with a global user base. 

Plato’s growth mirrored a broader trend. LinkedIn’s Workplace Learning Report shows that mentorship and coaching are among the fastest-growing professional development investments, driven by distributed work and leadership shortages in technical fields.

“Mentorship is infrastructure,” Salmon says. “You don’t see it until it’s missing.”

Later, at Wave.ai, he worked on AI-powered coaching systems - tools designed not to score or rank people, but to support reflection and growth. It was here that Salmon became increasingly vocal about the limitations of purely performance-driven AI.

“Most AI is optimised for output,” he says. “But humans don’t work that way. We need context, continuity, and emotional bandwidth.”

Similar concerns have been documented by Stanford’s Human-Centered AI Institute, which reports that user trust and adoption drop sharply when AI systems prioritise efficiency over interpretability and empathy - particularly in high-stakes or interpersonal domains.

Why the Future of AI Is Quietly Personal

In 2024, Salmon founded Super-me, a messaging and client-management platform for coaches, therapists, and solopreneurs. Built end-to-end by Salmon as a solo founder, the product focused on a narrow use case for client communication and management. 

“It wasn’t about building the biggest platform,” he says. “It was about proving that AI could support human relationships without flattening them.”

The timing was significant. The World Bank estimates that more than 500 million people globally now work as independent professionals, many in roles where trust and communication are core to economic value. In such roles, automation that supports communication and organisation may be more useful than tools designed to replace human interaction. 

Super-me was an experiment in that direction, and it informed Salmon’s current work as a founding engineer at Vybe, where he is helping build AI systems that allow non-engineers to create internal tools safely and at scale.

The economic implications are substantial. Internal software tooling is estimated by Gartner to account for a significant share of enterprise development spend, often bottlenecked by limited engineering resources. By reducing the dependency on custom engineering for internal systems, platforms like the one Salmon is building aim to reduce bottlenecks in internal tool development. 

“Engineers shouldn’t be gatekeepers,” he says. “They should be enablers.”

What distinguishes Salmon’s contribution is not just technical execution, but restraint. In his widely read essays, he argues that AI should industrialise the repetitive parts of software creation while preserving human judgement as the final authority.

“Anyone can make a system sound intelligent,” he says. “The real question is whether it helps people think more clearly.”

That ethos has informed his writing and discussions within engineering communities. . His work does not chase spectacle; it focuses on reliability, trust, and long-term usefulness - qualities that matter deeply in both economic systems and immigration frameworks that assess sustained contribution.

As nations compete for experienced technical leaders who can translate advanced technology into real-world impact, profiles like Salmon’s stand out. He represents a class of global professionals whose value lies not only in what they build, but in how they build it - and who they enable along the way.

“The future of technology won’t be louder,” he says. “It’ll be calmer. And more human.”